Predicting Vehicles' Longitudinal Trajectories and Lane Changes on Highway On-Ramps

TL;DR

This paper introduces a prediction framework for vehicle trajectories and lane changes on highway on-ramps, aiming to reduce congestion and enhance safety by forecasting vehicle behavior up to 15 seconds ahead.

Contribution

The paper presents a novel prediction framework combining models that outperform traditional LSTM models in forecasting vehicle trajectories and lane changes on on-ramps.

Findings

Prediction accuracy exceeds traditional LSTM models.

Framework effectively forecasts 15 seconds ahead.

Potential to improve traffic management and safety.

Abstract

Vehicles on highway on-ramps are one of the leading contributors to congestion. In this paper, we propose a prediction framework that predicts the longitudinal trajectories and lane changes (LCs) of vehicles on highway on-ramps and tapers. Specifically, our framework adopts a combination of prediction models that inputs a 4 seconds duration of a trajectory to output a forecast of the longitudinal trajectories and LCs up to 15 seconds ahead. Training and Validation based on next generation simulation (NGSIM) data show that the prediction power of the developed model and its accuracy outperforms a traditional long-short term memory (LSTM) model. Ultimately, the work presented here can alleviate the congestion experienced on on-ramps, improve safety, and guide effective traffic control strategies.